2008-09-27 |
Kevin Brubeck Unhammer | Update to remove faulty gold parses. Seems Done now :)
|
commit | commitdiff | tree |
2008-09-21 |
Kevin Brubeck Unhammer | minor additions
|
commit | commitdiff | tree |
2008-09-21 |
Kevin Brubeck Unhammer | report almost done
|
commit | commitdiff | tree |
2008-09-18 |
Kevin Brubeck Unhammer | another little fix to stop initialization
|
commit | commitdiff | tree |
2008-09-18 |
Kevin Brubeck Unhammer | another little fix to stop initialization
|
commit | commitdiff | tree |
2008-09-18 |
Kevin Brubeck Unhammer | rewrote loc_h_harmonic's STOP initialization to reflect...
|
commit | commitdiff | tree |
2008-09-18 |
Kevin Brubeck Unhammer | fixed a little bug in root reestimation (forgot to...
|
commit | commitdiff | tree |
2008-09-03 |
Kevin Brubeck Unhammer | before removing sum_hat_a from p_ATTACH
|
commit | commitdiff | tree |
2008-09-03 |
Kevin Brubeck Unhammer | running manual EM on sentence NVN
|
commit | commitdiff | tree |
2008-08-29 |
Kevin Brubeck Unhammer | started on pure cnf version, see sec6 of formulas.pdf
|
commit | commitdiff | tree |
2008-08-27 |
Kevin Brubeck Unhammer | loc_h_dmv reestimation conforms to formulas.pdf, but...
|
commit | commitdiff | tree |
2008-08-20 |
Kevin Brubeck Unhammer | first attempt at fixing the w-bug, todo: test
|
commit | commitdiff | tree |
2008-08-20 |
Kevin Brubeck Unhammer | found a bug in w-formula, 1/P_s should be outside sums...
|
commit | commitdiff | tree |
2008-08-11 |
Kevin Brubeck Unhammer | why is cnf-reestimate so slow??
|
commit | commitdiff | tree |
2008-08-10 |
Kevin Brubeck Unhammer | error in main.py with cnf reestimation, todo
|
commit | commitdiff | tree |
2008-08-10 |
Kevin Brubeck Unhammer | written reest for cnf, todo: test it
|
commit | commitdiff | tree |
2008-08-10 |
Kevin Brubeck Unhammer | texed some reestimation stuff for cnf_
|
commit | commitdiff | tree |
2008-08-10 |
Kevin Brubeck Unhammer | inner and outer done for cnf_dmv.py, seems to work...
|
commit | commitdiff | tree |
2008-08-01 |
Kevin Brubeck Unhammer | it seems cnf_dmv.inner() <=> loc_h_dmv.inner()... where...
|
commit | commitdiff | tree |
2008-07-30 |
Kevin Brubeck Unhammer | working on cnf_ files now, still lots to do but have...
|
commit | commitdiff | tree |
2008-07-27 |
Kevin Brubeck Unhammer | add_root fn written
|
commit | commitdiff | tree |
2008-07-27 |
Kevin Brubeck Unhammer | ...
|
commit | commitdiff | tree |
2008-07-27 |
Kevin Brubeck Unhammer | ...
|
commit | commitdiff | tree |
2008-07-27 |
Kevin Brubeck Unhammer | some files I forgot to add
|
commit | commitdiff | tree |
2008-07-27 |
Kevin Brubeck Unhammer | loc_h_dmv IO and reest done, need to do main.py:def...
|
commit | commitdiff | tree |
2008-07-08 |
Kevin Brubeck Unhammer | Rewrote inner and outer to match klein-thesis
|
commit | commitdiff | tree |
2008-07-05 |
Kevin Brubeck Unhammer | still working on between-words
|
commit | commitdiff | tree |
2008-07-05 |
Kevin Brubeck Unhammer | trying between-words
|
commit | commitdiff | tree |
2008-06-23 |
Kevin Brubeck Unhammer | before trying new pchoose reestimation
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | PSTOP done, PCHOOSE close to done--but still gives...
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | ...
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | close to finishing pCHOOSE
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | 100x speed increase: yay! but how to do pCHOOSE?
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | pstop seems done, todo: pchoose reestimation
|
commit | commitdiff | tree |
2008-06-15 |
Kevin Brubeck Unhammer | pstop seems done, todo: pchoose reestimation
|
commit | commitdiff | tree |
2008-06-12 |
Kevin Brubeck Unhammer | c(s,t,i) done now
|
commit | commitdiff | tree |
2008-06-12 |
Kevin Brubeck Unhammer | outer seems to be working now
|
commit | commitdiff | tree |
2008-06-11 |
Kevin Brubeck Unhammer | some fixes to outer
|
commit | commitdiff | tree |
2008-06-11 |
Kevin Brubeck Unhammer | todo: test outer() probabilites
|
commit | commitdiff | tree |
2008-06-10 |
Kevin Brubeck Unhammer | inner now makes trees as well, for pruning
|
commit | commitdiff | tree |
2008-06-10 |
Kevin Brubeck Unhammer | some optimizations, inner is about 3 times faster now
|
commit | commitdiff | tree |
2008-06-09 |
Kevin Brubeck Unhammer | todo: test PSTOP(h|ln)
|
commit | commitdiff | tree |
2008-06-09 |
Kevin Brubeck Unhammer | todo:general reestimation
|
commit | commitdiff | tree |
2008-06-08 |
Kevin Brubeck Unhammer | made a pruning function for those unused subtrees....
|
commit | commitdiff | tree |
2008-06-08 |
Kevin Brubeck Unhammer | a bit better, but still not sure P_STOP does everything...
|
commit | commitdiff | tree |
2008-06-08 |
Kevin Brubeck Unhammer | Somehow must get counts only from _completed_ trees...
|
commit | commitdiff | tree |
2008-06-07 |
Kevin Brubeck Unhammer | Ok, so I misunderstood the word adjacent all along...
|
commit | commitdiff | tree |
2008-06-07 |
Kevin Brubeck Unhammer | todo: fix P_STOP so it actually changes stuff
|
commit | commitdiff | tree |
2008-06-07 |
Kevin Brubeck Unhammer | inner_dmv should be working again--this time with loc_h...
|
commit | commitdiff | tree |
2008-06-06 |
Kevin Brubeck Unhammer | inner_dmv() rewritten with loc_h, works with simple...
|
commit | commitdiff | tree |
2008-06-06 |
Kevin Brubeck Unhammer | new version of inner_dmv and rule.p written, must test
|
commit | commitdiff | tree |
2008-06-06 |
Kevin Brubeck Unhammer | sthg wrong with P_STOP
|
commit | commitdiff | tree |
2008-06-06 |
Kevin Brubeck Unhammer | started working on new version of inner
|
commit | commitdiff | tree |
2008-06-05 |
Kevin Brubeck Unhammer | Usable test of PSTOP reestimation
|
commit | commitdiff | tree |
2008-06-05 |
Kevin Brubeck Unhammer | still working on pstop
|
commit | commitdiff | tree |
2008-06-05 |
Kevin Brubeck Unhammer | more backups
|
commit | commitdiff | tree |
2008-06-05 |
Kevin Brubeck Unhammer | backups
|
commit | commitdiff | tree |